An altogether fresh approach to behavior has exposed an unsuspected dimension of alcohol. We have identified behavioral variation (BV) as a fundamental aspect of behavior governed by its own neural mechanisms and having its own adaptive value. Although it participates in the commission of errors, it also intelligently arranges for behavioral adjustment to labile environments. Much preliminary work has shown that alcohol selectively narrows BV. Thus, alcohol shares an unsuspected attribute of other abused durgs: it promotes stereotypy. However, it is not stereotypy per se that is of interest to us. Alcohol attenuates the production of variety: it impairs the otherwise enforced exposure to new places, topographies, rates, etc. Stereotypy does not appear as an emergent pattern -- it is what remains when BV is lost. Our concern, therefore, is with the ordinary adaptive value of BV -- its role in providing for the discovery of new contingencies or adapting economically to the loss of old ones. An understanding of this adaptive function, and its opposition by ethanol has permitted the unification of much that is conflicting or anomalous in the learning-and-alcohol literature. Alcohol, as we have found and propose to investigate in detail, frequently improves performance in some situations and impairs it in others. We predict that alcohol will act synergistically with the performance of tasks whose mastery entails invariance, and antagonistically with tasks requiring variation. Most importantly, it is expected to impair adjustment to nearly any shift in task requirement. Experimental settings of these sorts are drawn from procedrues that emphasize temporal (operant) and spatial (maze) contingencies. These experiments are intended to test our hypotheses about alcohol and instrumental performance. But of equal importance, they permit the direct measurement of BV in parallel with conventional performance data, allowing us to draw correspondences or dependencies between the two sets of variables. Our preliminary work has strongly implicated the frontal cortex and the hippocampus in the regulation of BV. Accordingly, we propose that these are important sites of alcohol action. We test the possibility that the effect of ethanol on BV is mediated through these structures.